DSR²

A modern and fast scientific paper search engine built from scratch with Python.

Features

• Lightning Fast: Sub-second search results with optimized indexing
• Intelligent Ranking: Modern relevance scoring algorithms (TF-IDF)
• Synonym search: Fast query expansion with synonyms, powered by machine learning
• Additional Documents: Allow the user to upload documents and update indexes near-instantly
• Modern UI: Clean, responsive interface built with React
• Mobile Ready: Fully responsive design for all devices

Tech Stack

Performance

• Search Speed: < 1s average response time
• Corpus: 1 million+ scientific and academic papers

Query Processing

📋 Project Structure

DSR²/
├── Assets/
├── backend/
├── Barreling/
│   ├── BarrelingInverted.py
│   ├── Barrels.ipynb
│   └── actualBarrels.py
├── Cleaning/
│   └── DataCleaning.py
├── Indexing/
│   ├── Forward_Index.py
│   ├── Inverted_Index.py
│   └── lexicon.py
├── Searching/
│   ├── app.py
│   ├── documentAddition.py
│   ├── documentRetrieval.py
│   ├── ranking.py
│   ├── utils.py
│   └── requirements.txt
└── frontend/
    ├── app/
    ├── components/
    ├── lib/
    ├── public/
    └── styles/

Data

The datasets and indices can all be downloaded here.

License

This project is licensed under the MIT License - see the LICENSE file for details.

Acknowledgments

• Kaggle for the large scientific paper dataset
• Flask for the excellent Python framework
• React for the amazing frontend library
• v0.dev for allowing us to build such an amazing frontend

---

Made with ❤️ by

Danish Munib, Nouman Ghauri and Muhammad Anas